Immunogenicity Publication Bias and Its Consequences for Predictive Models: A Call for Transparent Reporting.

Area of Science:

• Biopharmaceutical development
• Immunology
• Clinical trial analysis

Background:

• Therapeutic protein immunogenicity, specifically antidrug antibody (ADA) development, is a critical factor in drug efficacy and safety.
• Accurate prediction of ADA incidence is essential for successful drug development and regulatory approval.

Purpose of the Study:

• To investigate the bias in published antidrug antibody (ADA) incidence data for therapeutic proteins.
• To analyze the impact of this bias on the prediction of ADA incidence over time.
• To emphasize the importance of transparent data reporting across the biopharmaceutical industry.

Main Methods:

• Comparative analysis of Phase I-III ADA incidence data for internal and approved monoclonal antibodies.
• Development of an empirical model to assess the effect of data bias on ADA incidence time-course predictions.

Main Results:

• Published ADA incidence data exhibit a bias towards lower reported rates compared to comprehensive trial data.
• Omission of early-phase trial data, particularly from trials discontinued due to high immunogenicity, significantly contributes to this bias.
• The identified bias leads to inaccurate predictions of ADA incidence over time.

Conclusions:

• The current reporting of ADA incidence data is likely an underestimation of the true immunogenicity risk associated with therapeutic proteins.
• Transparent and complete data sharing from all clinical trial phases is crucial for accurate risk assessment and improved therapeutic protein development.
• Addressing this reporting bias is essential for enhancing the predictability of immunogenicity and optimizing the development of safe and effective protein-based therapeutics.